MR. W. CROOKES OK REPULSION RESULTING FROM RADIATION. 
97 
I was unable to produce negative rotation, or a position of rest, with this form of 
instrument. 
Fig. 10. 
404. My next endeavour was to ascertain what difference in action was caused by 
an alteration in the size and shape of the screens. A radiometer was constructed like 
the one shown in fig. 8 (400), except that the mica disks which acted as screens were 
2 millims. smaller in diameter than the metal cups. The screens were adjustable. 
This instrument was found to be less sensitive than the one described in par. 400, 
where the screens and cups were of the same diameter, and under the influence of the 
candle it behaved differently. When the screens were close to the concave side of the 
cup, as shown at G, fig. 8, the revolutions were 20 per minute. When they were 
midway between the cups, as at E, fig. 8, the revolutions were 10 per minute. With 
the screens close to the convex sides, as at A, fig. 8, there was no movement at all. 
When rotation occurred it was in each case positive. The rotation when the screens 
were 2 millims. from the convex sides of the cups, as at B, fig. 8, was positive at the 
rate of 3 revolutions a minute. With the larger screens, as described and figured in 
par. 400, the rotation in this position was negative, at the rate of 3 per minute. 
The exhaustion in these experiments was 12 M. 
405. Other radiometers were made of the forms shown in fig. 11. In A, the mica 
Fig. 11. 
35 revolutions a minute. 
screen was absent; in B it was in the form of a half disk covering the inner half 
of the cup, and in C a similar screen covered the outer half of the cup. The form A 
was the least sensitive, and the form C most so ; the speed in forms A, B, and C 
being 35, 66, and 100 revolutions a minute. 
MDCCCLXXJX. 
o 
